JPH02153478A - Individual collating device - Google Patents

Abstract

PURPOSE:To speed up a line thinning processing by detecting the position of a sweat gland in a fingerprint image, selecting the sweat gland belonging onto the same raised line, and connecting each sweat gland belonging onto the same raised line by a segment. CONSTITUTION:A device is provided with each function part of a fingerprint sensor 1, a gradation image storage device 2, a binarization circuit 3, a binarization image storage device 4, a sweat gland detecting circuit (seat gland detecting means) 5, an investigating circuit (selecting means) 6, a coordinate management directory 7 of a sweat gland, a combination managing means 8 of a sweat gland, an H-shaped structure detecting circuit 9, a feature extracting circuit (connecting means) 10, and a registering/collating circuit 11. Accordingly, each sweat gland belonging to the same raised line is connected by a segment, any by this segment, a fine line-converted image is made. In such a manner, it becomes unnecessary to execute the line thinning processing by a picture element unit, and the processing can be executed at high speed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a personal identification device, and more particularly to a personal identification device that identifies individuals using fingerprints.

In recent years, as computers have become widespread throughout society, public interest has focused on how to ensure their integrity (security). In the past, ID cards and PINs were commonly used to verify identity when entering a computer room or using a terminal. However, once the magnetic information on an ID card is duplicated, it is easy to be infiltrated using a counterfeit card. Furthermore, the PIN number is usually an easy-to-remember date of birth, telephone number, etc., and can be easily discovered by checking information around the person.

Fingerprints, on the other hand, are considered the most powerful means of identifying individuals because they have three major properties: ``university for all people'' and ``unchangeable throughout life.'' Many studies have been conducted on simple personal identification systems using fingerprints. Development is underway.

[Conventional technology]

In the development of a personal verification device using fingerprints, we aim to realize a real-time fingerprint sensor that can clearly input the fingerprint image of the person in real time, and to combine the registered characteristic information and the input fingerprint with sufficient information for practical use. The key is to develop algorithms that perform high-speed matching while maintaining high reliability.

In a real-time fingerprint sensor, it is necessary to input a clear, distortion-free fingerprint image without being affected by the residual fingerprint of the previous user. The present inventor has developed a holographic fingerprint sensor with a new configuration that combines a light plate.

On the other hand, when developing algorithms, it is necessary to flexibly deal with differences in the way force is applied during input and distortion of the fingerprint image caused by moving the finger while it is on the glass. The present inventor has newly developed a moving window algorithm that combines the redundancy of the pattern matching method and the structural analysis of the feature extraction method.

Generally, when an image mainly consisting of lines (ridges and valleys), such as a fingerprint, is to be binarized, information regarding the thickness of the lines is not so important. Therefore, when converting a fingerprint into a binary figure, a so-called line thinning process is performed, which is a method of extracting the center line of a line and using the extracted line to represent the ridges and valleys of the fingerprint. Thinning processing is also effective as a method for compressing information and contributes to high-speed image processing. As an example of such thinning processing, the outermost pixels of the ridge area are repeatedly deleted until the inner area becomes empty, and the series of pixels that remain without being deleted is used as the center line. Specifically, from an arbitrary figure area Do in a given binary figure, delete the pixels that constitute its outline but are not multipoint points, and let the resulting figure be Dl.Similarly, , from Dl to D2, from D2 to D3, etc., are repeated until there are no more pixels to delete.Then, if the finally obtained figure is Dn, then Dn is DO The center line of

[Problem to be solved by the invention]

However, such conventional personal verification devices are configured to determine whether each pixel is deleted or not deleted in the line thinning process, so fingerprint patterns are generally complex and it is difficult to distinguish between them. Since the number of pixels displayed is large, there is a problem in that the processing speed takes a long time.

Therefore, an object of the present invention is to obtain a thinned image without using a complicated algorithm such as pixel-by-pixel deletion or non-deletion, and to speed up image processing, particularly thinning processing.

[Means to solve the problem]

In order to achieve the above object, the personal verification device according to the present invention is a personal verification device that thins an input fingerprint image and registers and verifies the fingerprint of a specific individual based on the thinning data. Sweat gland detection means for detecting the position of sweat glands in the fingerprint image, selection means for selecting sweat glands belonging to the same ridge, and connecting sweat glands belonging to the same ridge selected by the selection means with line segments. and a connecting means.

[Effect]

In the present invention, sweat glands belonging to the same ridge are connected by line segments, and a thinned image is created by the line segments. Therefore, there is no need to perform thinning processing on a pixel by pixel basis, and processing speed can be increased.

〔Example〕

Hereinafter, the present invention will be explained based on the drawings.

1 to 6 are diagrams showing an embodiment of the personal verification device according to the present invention.

First, the configuration will be explained. FIG. 1 is a block diagram showing the functions of the personal verification device. In this diagram, the personal verification device includes, for example, a 16-bit personal computer, and is roughly divided into a fingerprint sensor 1, a grayscale image storage device 2, Binarization circuit 3, binary image storage device 4, sweat gland detection circuit (sweat gland detection means) 5, investigation circuit (selection means) 6, sweat gland coordinate management book 7, sweat gland combination management book 8, H-shaped It has functional units represented by a structure detection circuit 9, a feature extraction circuit (connection means) 10, and a registration/verification circuit 11.

The fingerprint sensor 1 is, for example, a type that uses a laser as a light source, a combination of a hologram and a flat light guide plate, or a prism type, and receives a fingerprint with an image sensor and converts it into an image signal. The grayscale image storage device 2 stores an image by A/D converting the image signal from the fingerprint sensor 1, and the stored data is sent to the binarization circuit 3 and binarized.

In addition, in the above binarized data, the ridges (convex and mountainous parts) of the fingerprint are "1", and the valley lines (concave and valley parts) are "1".
It is a so-called silhouetted binary figure corresponding to 0". The binary image storage device 4 stores the above-mentioned binary data, and the data stored in the binary image storage device 4 is sent to the sweat gland detection circuit 5. At the same time, it is also sent to an investigation circuit 6 that investigates pixels on line segments between sweat glands.The sweat gland detection circuit 5 detects sweat glands from the binary data sent from the binary image storage device 4. It detects the position, and the detection method is as follows.
For example, you can find a branching point of a contour line in the binarized data, draw an inspection line perpendicular to the straight line connecting any two branching points, and if this inspection line If it intersects with the sweat gland, you can use the method of treating it as a sweat gland.

Incidentally, in this method, the contour line of the sweat gland always constitutes a closed curve,
However, it is based on the principle that the size is a general rule (
. The sweat gland coordinate data detected by the sweat gland detection circuit 5 is sent to a sweat gland coordinate management book 7, and this coordinate management book 7 stores and manages a large number of sent sweat gland coordinate data. The investigation circuit 6 extracts one sweat gland coordinate A from the management data on the sweat gland coordinate management 787, that is, from a large number of sweat gland coordinates, and extracts one sweat gland coordinate A from the sweat gland coordinate management 787.
With reference to other sweat gland coordinates C, D, E,...
Investigate the pixels between... Specifically, as shown in Fig. 2, within a survey area of ±iV pixels square with A as the center, line segments ABSAC, Assuming AD,, AE, the coordinates of point C (see FIG. 3) on each line segment are calculated according to the following equation (2).

c=p-a+ (1-p) ・b ・Old・・■(0<
p<l) However, aSbSc is the position vector of sweat glands A, B, and point C when the line segment is AB.

In the above equation (2), if the parameter P is changed appropriately, C changes. That is, by changing as shown in Figure 3,
If we examine the pixel at point C, for example, AB shown in Figure 2,
In the case of AD, since the pixels are consecutive "°1"°, it is investigated that sweat glands A and B and A and D belong to the same ridge.

On the other hand, as for AC and AE, since they straddle the valley line, their pixels become "0"' in the valley line part, and it was investigated that these sweat glands A and C, and A and E do not belong to the same ridge. be done. The investigation results of the investigation circuit 6, ie, the data on the combinations of sweat glands belonging to the same ridge, are sent to the sweat gland combination management record 8, and are stored and managed in the sweat gland combination management record 8. The H-shaped structure detection circuit 9 removes unnecessary data due to "bridges" from the above combination data. "Bridges" are generated when valley lines are filled with moisture, etc., and this bridge is as if it were a ridge. Since it is misidentified, it must be completely removed from the data. In this embodiment, for example, in the case of a bridge as shown in FIG.
There is no line segment that passes through this bridge and connects any two sweat glands consecutively with a pixel of "1".Therefore, in the combination data managed in the sweat gland combination management record 8, as shown in FIG. Data for bridges such as
), if a bridge exists between sweat glands AB,
In this case, a line segment AB is drawn between the sweat glands AB, and since AB is continuous at a pixel of "°1", there is a risk that it may be mistaken for a ridge. However, the ridges caused by the bridge can be removed by providing the following conditions. In other words, if three or more line segments are drawn from a given sweat gland, 3 or more line segments are drawn from any other sweat gland that shares a line segment with this given sweat gland.
Examine the sweat glands that have three or more line segments drawn on them. Then, the line segment between this sweat gland and the above-mentioned arbitrary sweat gland may be removed as an erroneous line segment created by a "bridge." As a result, the fifth
For example, even if there is a bridge between the sweat glands AB as shown in Figure (b), the ridges due to this bridge can be removed. The feature extraction circuit 10 receives combination data of sweat glands belonging to the same ridge managed in the sweat gland combination management record 8, and generates a thinned image by connecting these combined sweat glands with line segments. , feature points (for example, Y-branch points and end points) are extracted from this thinned image. The registration/verification circuit 11 registers the extracted feature points as verification data representing the fingerprint of a specific individual, and at the time of verification, performs a verification operation for the specific individual using this verification data.

Next, the effect will be explained.

FIG. 6 is a flowchart showing a fingerprint registration program in the personal verification process, and P+ to PIO in the figure are step numbers of the program.

First, a fingerprint image of a specific individual to be registered is input using P. In this embodiment, the image size is, for example, 256.
256 pixels is used. Next, the input image is binarized in P2. In this binarization, for example, the image is
Dividing the image into small areas of 6 x 16 pixels, finding the average density of the grayscale image for each area, and using that value as a threshold value for binarization, can deal with the grayscale distribution of the image due to imbalances in finger pressure during input, etc. This is preferable because it can be done.

Next, P detects a large number of sweat gland coordinates from the binarized data and registers them in sweat gland coordinate management'; IJ7. P
In steps 4 to P7, the pixels on the line segment connecting one sweat gland and other sweat glands in the vicinity are checked, and if "1" is consecutive, the combination of sweat glands is considered to be on the same ridge, and 0" If there are any combinations of sweat glands that cross the valley line and are on different ridges, we investigate the process of removing that combination for all sweat glands.Next, in P,~PIG, for all sweat gland combinations, In other words, it is determined whether an I (shaped structure exists or not, and if so, that combination of sweat glands is removed.Finally, in Pl+, based on the sweat gland combination data that does not include a bridge) Feature points are extracted and registered from the thinned image created, and matching operations are performed using these feature points as necessary.

As described above, according to this embodiment, a thinned image is created by connecting sweat glands that belong to the same ridge. Therefore, it is not necessary to use a complicated algorithm for determining deletion or non-deletion for each of a large number of pixels forming a ridge figure, as in the past, and processing speed can be increased. Furthermore, in this embodiment, the bridge can also be easily removed.

(Effects of the Invention) According to the present invention, a thinned image can be obtained without using a complicated algorithm such as deletion/non-deletion for each pixel.
It is possible to speed up image processing.

[Brief explanation of the drawing]

1 to 6 are diagrams showing one embodiment of the personal verification device according to the present invention, FIG. 1 is a functional block diagram thereof, FIG. 2 is a conceptual diagram for investigating the relationship between sweat glands, and FIG. 3 is a diagram explaining an arbitrary point C between the sweat glands AB, Figure 4 is a diagram when the sweat gland and the bridge are far apart, and Figures 5 (a) and (b) are diagrams when the sweat gland and the bridge are close. Figure 6 is a flowchart of the fingerprint registration program. 5...Sweat gland detection circuit (sweat gland detection means), 6...
... Investigation circuit (selection means), 10 ... Feature extraction circuit (connection means) Conceptual diagram for investigating the relationship between sweat glands of one embodiment Second factor diagram Diagram sweat glands of one embodiment Figure 4 - 1m91oi''5

Claims (1)

[Scope of Claims] In a personal verification device that performs thinning processing on an input fingerprint image and registers and verifies the fingerprint of a specific individual based on the thinning data, the position of a sweat gland in the fingerprint image is detected. Sweat gland detection means; selection means for selecting sweat glands belonging to the same ridge; and connection means for connecting the sweat glands belonging to the same ridge selected by the selection means with line segments. Personal verification device.